Application of an in vitro new approach methodology to determine relative cancer potency factors of air pollutants based on whole mixtures.

Air pollution is an example of a complex environmental mixture with different biological activities, making risk assessment challenging. Current cancer risk assessment strategies that focus on individual pollutants may overlook interactions among them, potentially underestimating health risks. Therefore, a shift towards the evaluation of whole mixtures is essential for accurate risk assessment. This study presents the application of an in vitro New Approach Methodology (NAM) to estimate relative cancer potency factors of whole mixtures, with a focus on organic pollutants associated with air particulate matter (PM). Using concentration-dependent activation of the DNA damage-signaling protein checkpoint kinase 1 (pChk1) as a readout, we compared two modeling approaches, the Hill equation and the benchmark dose (BMD) method, to derive Mixture Potency Factors (MPFs). MPFs were determined for five PM2.5 samples covering sites with different land uses and our historical pChk1 data for PM10 samples and Standard Reference Materials. Our results showed a concentration-dependent increase in pChk1 by all samples and a higher potency compared to the reference compound benzo[a]pyrene. The MPFs derived from the Hill equation ranged from 128 to 9793, while those from BMD modeling ranged from 70 to 303. Despite the differences in magnitude, a consistency in the relative order of potencies was observed. Notably, PM2.5 samples from sites strongly impacted by biomass burning had the highest MPFs. Although discrepancies were observed between the two modeling approaches for whole mixture samples, relative potency factors for individual PAHs were more consistent. We conclude that differences in the shape of the concentration-response curves and how MPFs are derived explain the observed differences in model agreement for complex mixtures and individual PAHs. This research contributes to the advancement of predictive toxicology and highlights the feasibility of transitioning from assessing individual agents to whole mixture assessment for accurate cancer risk assessment and public health protection.

High throughput screening of bisphenols and their mixtures under conditions of low-intensity adipogenesis of human mesenchymal stem cells (hMSCs).

In vitro models of adipogenesis are phenotypic assays that most closely mimic the increase of adipose tissue in obesity. Current models, however, often lack throughput and sensitivity and even report conflicting data regarding adipogenic potencies of many chemicals. Here, we describe a ten-day long adipogenesis model using high content analysis readouts for adipocyte number, size, and lipid content on primary human mesenchymal stem cells (MSC) sensitive enough to compare bisphenol A derivatives quantitatively in a robust and high throughput manner. The number of adipocytes was the most sensitive endpoint capable of detecting changes of 20% and was used to develop a benchmark concentration model (BMC) to quantitatively compare eight bisphenols (tested at 0.1-100 µM). The model was applied to evaluate mixtures of bisphenols obtaining the first experimental evidence of their additive effect on human MSC adipogenesis. Using the relative potency factors (RPFs), we show how a mixture of bisphenols at their sub-active concentrations induces a significant adipogenic effect due to its additive nature. The final active concentrations of bisphenols in tested mixtures reached below 1 µM, which is within the concentration range observed in humans. These results point to the need to consider the toxicity of chemical mixtures.

Benchmark dose-response analyses for multiple endpoints in drug safety evaluation.

Hazard characterization during pharmaceutical development identifies the candidate drug's potential hazards and dose-response relationships. To date, the no-observed-adverse-effect-level (NOAEL) approach has been employed to identify the highest dose which results in no observed adverse effects. The benchmark dose (BMD) modeling approach describes potential dose-response relationships and has been used in diverse regulatory domains, but its applicability for pharmaceutical development has not previously been examined. Thus, we applied BMD-modeling to all endpoints in three sequential in vivo studies in a drug development setting, including biochemistry, hematology, organ pathology and clinical observations. In order to compare the results across such a broad range of effects, we needed to standardize the choice of the critical effect size (CES) for the different endpoints. A CES of 5%, previously suggested by the European Food Safety Authority, was compared with the study NOAEL and with the General Theory of Effect Size, which takes natural variability into account. Compared to the NOAEL approach, the BMD-modeling approach resulted in more informative estimates of the doses leading to effects. The BMD-modeling approach handled well situations where effects occurred below the lowest tested dose and the study's NOAEL, and seems advantageous to characterize the potential toxicity during safety assessment. The results imply a considerable step forward from the perspective of reducing and refining animal experiments, as more information is yielded from the same number of animals and at lower doses. Taken together, employing BMD-modeling as a substitute, or as a complement, to the NOAEL approach seems appropriate.